1. Field of the Invention
The present general inventive concept relates to an apparatus and a method of detecting an error in a data storage device, and more particularly, to an apparatus and a method of detecting an error symbol to improve an error correction performance of an error correcting decoder.
2. Description of the Related Art
Japanese Laid-open Patent No. 2000-243034 and Korean Laid-open Patent No. 1999-061966 relates to an equalizing process of a disk storage device.
In Japanese Laid-open Patent No. 2000-243034, an equalization processing technology, by which a symmetric finite impulse response (FIR) filter and an asymmetric FIR filter are connected to an output of a sampler in parallel, a phase of a phase locked loop (PLL) circuit is adjusted only using an output of the symmetric FIR filter, and an equalization is processed using the asymmetric FIR filter. In addition, Korean Laid-open Patent No. 1999-061966 discloses an equalization processing technology that is able to reduce error signals efficiently even when a signal obtained from a channel is asymmetric.
In general, hard disk drives that are data storage devices contribute to operate computer systems by reproducing data recorded in disks using a magnetic head or by recording data in the disk. The hard disk drives are being formed to have larger capacity, higher density and a small size. Accordingly, a bit per inch (BPI) that is a recording density in a rotary direction of the disk and a track per inch (TPI) that is a recording density in a diameter direction of the disk increase, and accordingly, a precise mechanism is required.
A data detecting device used in the hard disk drive currently is designed based on a partial response maximum likelihood (PRML) detection technology. The PRML detection device may increase a bit error rate (BER) due to media noises generated in recorded data patterns.
In order to compensate for the above problem, a post processor is added to the data detecting device to reduce the BER. However, the conventional post processor can detect the error only when an error of a maximum of 1 bit is generated at every codeword.
The conventional post-processing has at least two problems.
First, if the location of error generation is incorrectly determined and the supposed error is corrected, originally correct data is changed as if it is the error data, and thus, the BER increases.
Second, when the recording density of the hard disk drive increases, a signal to noise ratio (SNR) is reduced, and thus, two or more bits of errors may be generated per codeword. Therefore, the error cannot be corrected by the post processor.